Combite MPRNet (derain) + YOLO11s

Not dont yet, can only Derain

still working on connect with USB camera

Installation

Use the package manager pip to install dependecies.

conda create --name venv_mprnet python=3.8
conda activate venv_mprnet
pip install numpy ultralytics
pip install matplotlib scikit-image opencv-python yacs joblib natsort h5py tqdm

Example to test code library

python demo_yolo.py --task Deraining --input_dir ./Images/Raw_images/ --result_dir ./Images/Result_images/

Example code for run MPRNet_Yolo_ver1.py

For now, can ONLY Derain, read image from folder, YOLO can drawing the rectangle prediction at result images, Pruning YOLO Model.

python MPRNet_Yolo_ver1.py --input_dir ./Images/Raw_images --result_dir ./Images/Result_images/ --task Deraining

Pruning MPRNet, connecting with USB Camera is underdevelopment . . .

Example code for run MPRNet_Yolo_Camera.py

Pipeline : USB Camera => MPRNet(pruned+quantized) + YOLO(pruned) => show detect image via opencv (not save to folder)

can choose a task 'deblurring', 'denoising', 'deraining'

can choose USB Camera Index e.g. 0, 1, 2

can choose image resolution e.g. 720x480, 480x320

python MPRNet_Yolo_Camera.py --task deraining --usb_cam_idx 0 --img_res 480x320

reference

https://github.com/swz30/MPRNet

note

Right now, cann't connect with usb camera yet. just scan image in folder. MPRNet and YOLO not connect each other yet.